Anaerobic fermentations of hydrogen and carbon monoxide involve the contact of the substrate gas in a liquid fermentation menstruum with microorganisms capable of generating oxygenated organic compounds such as ethanol, acetic acid, propanol and n-butanol. The production of these oxygenated organic compounds requires significant amounts of hydrogen and carbon monoxide. For instance, the theoretical equations for the conversion of carbon monoxide and hydrogen to ethanol are:6CO+3H2O→C2H5OH+4CO2 6H2+2CO2→C2H5OH+3H2O.
As can be seen, the conversion of carbon monoxide results in the generation of carbon dioxide. The conversion of hydrogen involves the consumption of hydrogen and carbon dioxide, and this conversion is sometimes referred to as the H2/CO2 conversion. For purposes herein, it is referred to as the hydrogen conversion.
Typically the substrate gas for carbon monoxide and hydrogen conversions is, or is derived from, a synthesis gas (syngas) from the gasification of carbonaceous materials, from the reforming of natural gas and/or biogas from anaerobic digestion or from off-gas streams of various industrial methods. The gas substrate contains carbon monoxide, hydrogen, and carbon dioxide and usually contains other components such as water vapor, nitrogen, methane, ammonia, hydrogen sulfide and the like. (For purposes herein, all gas compositions are reported on a dry basis unless otherwise stated or clear from the context.)
These substrate gases are typically more expensive than equivalent heat content amounts of fossil fuels. Hence, a desire exists to use these gases efficiently to make higher value products. The financial viability of any conversion process, especially to commodity chemicals such as ethanol and acetic acid, will be dependent upon capital costs, the efficiency of conversion of the carbon monoxide and hydrogen to the sought products and the energy costs to effect the conversion.
Syngas fermentation processes suffer from the poor solubility of the gas substrate, i.e., carbon dioxide and hydrogen, in the liquid phase of the fermentation menstruum where the biological processes occur. Munasinghe, et al., in Biomass-derived Syngas Fermentation in Biofuels: Opportunities and Challenges, Biosource Technology, 101 (2010) 5013-5022, summarize volumetric mass transfer coefficients to fermentation media that are reported in the literature for syngas and carbon monoxide in various reactor configurations and hydrodynamic conditions. A number of conditions can enhance the mass transfer of syngas to the liquid phase. For instance, increasing the interfacial area between the gas feed and the liquid phase can improve mass transfer rates.
Numerous processes have been disclosed for the conversion of carbon monoxide and hydrogen to oxygenated compounds. One such process suspends the microorganisms for the conversion in an aqueous menstruum contained in a stirred tank reactor such as by using a motor driven impeller. Stirred tank fermentation reactors provide many advantages. For stirred tank reactors, increasing the agitation of the impeller is said to improve mass transfer as smaller bubble sizes are obtained. Also, the stirring action not only distributes the gas phase in the aqueous menstruum but also the duration of the contact between the phases can be controlled. Another very significant benefit is that the composition within the stirred tank can be relatively uniform. For instance, Munasignhe, et al., in a later published paper, Syngas Fermentation to Biofuel: Evaluation of Carbon Monoxide Mass Transfer Coefficient (kLa) in Different Reactor Configurations, Biotechol. Prog., 2010, Vol. 26, No. 6, pp 1616-1621, combine a sparger (0.5 millimeter diameter pores) with mechanical mixing at various rotational rates to provide enhanced mass transfer. This uniformity enables good control of the fermentation process during steady-state operation. This is of particular advantage in the anaerobic conversion of carbon monoxide and hydrogen to oxygenated compounds where two conversion pathways exist. Hence the carbon dioxide generated from the conversion of carbon monoxide is proximate in location to the hydrogen consumption pathway that consumes carbon dioxide. The uniformity further facilitates the addition of fresh gas substrate. The problems with stirred tank reactors are capital costs, the significant amount of energy consumed in the needed mixing and agitation, and the need for plural stages to achieve high conversion of substrate.
Bredwell, et al., in Reactor Design Issues for Synthesis-Gas Fermentations, Biotechnol. Prog., 15 (1999) 834-844, disclose using microbubble sparging with mechanical agitation. At page 839 they state:
“When microbubble sparging is used, only enough power must be applied to the reactor to provide adequate liquid mixing. Thus axial flow impellers designed to have low shear and a high pumping capacity would be suitable when microbubbles are used in stirred tanks.”They conclude by stating:“An improved ability to predict and control coalescence rates is needed to rationally design commercial-scale bioreactors that employ microbubble sparging.” (p. 841)
Another type of fermentation apparatus is a bubble column fermentation reactor wherein the substrate gas is introduced at the bottom of the vessel and bubbles through the aqueous menstruum (“bubble reactor”). See Munasinghe, et al., in Biomass-derived Syngas Fermentation in Biofuels: Opportunities and Challenges, Biosource Technology, 101 (2010) 5013-5022. In order to achieve sought mass transfer from the gas to liquid phases, workers have provided the gas feed to bubble columns in the form of microbubbles. The authors report that in one study, the mass transfer obtained for a bubble column reactor was higher than that for a stirred tank reactor mainly due to the higher interfacial surface area obtained with the bubble column reactor. Advantageously, commercial-scale bubble column fermentation reactors are relatively simple in design and construction and require relatively little energy to operate.
In co-pending U.S. patent application No. 13/243,426, filed on even date herewith, processes are disclosed for enhancing the performance of large-scale, anaerobic fermentors. In these processes, a reactor having an aqueous menstruum depth of at least about 10 meters is used, and gas feed is supplied to the aqueous menstruum in the form of a stable gas-in-liquid dispersion. The aqueous menstruum is mechanically stirred at a rate sufficient to provide relatively uniform liquid phase composition within the aqueous menstruum without unduly adversely affecting the gas-in-liquid dispersion. For purposes herein, this is referred to as a mechanically-assisted liquid distribution tank reactor, or MLD tank reactor. At least a portion of the off-gas from the aqueous menstruum is recycled to obtain a molar conversion efficiency of total hydrogen and carbon monoxide in the gas substrate to oxygenated organic compound of at least about 80 percent in a single reactor stage. Accordingly, capital cost savings and energy savings over a conventional stirred tank reactor can be obtained.
For purposes herein, both deep, bubble column fermentation reactors and the large-scale MLD tank reactors supplied with stable gas feed-in-liquid dispersions and using low stifling rates, are referred to as deep tank reactors.
Deep tank reactors using microbubbles can provide economically attractive facilities for anaerobic conversion of syngas to oxygenated organic compound, but difficulties are present. In their earlier review article, Munasignhe, et al., report that the gas-liquid mass transfer is the major resistance for gaseous substrate diffusion. The authors state at page 5017:
“High pressure operation improves the solubility of the gas in the aqueous phase. However, at higher concentrations of gaseous substrates, especially CO, anaerobic microorganisms are inhibited.”
Other workers have understood that the presence of excess carbon monoxide can adversely affect the microorganisms and their performance. See paragraphs 0075 through 0077 and 0085 though 0086 of United States published patent application No. 20030211585 (Gaddy, et al.) disclosing a continuously stirred tank bioreactor for the production of ethanol from microbial fermentation. At paragraph 0077, Gaddy, et al., state:
“The presence of excess CO unfortunately also results in poor H2 conversion, which may not be economically favorable. The consequence of extended operation under substrate inhibition is poor H2 uptake. This eventually causes cell lysis and necessary restarting of the reactor. Where this method has an unintended result of CO substrate inhibition (the presence of too much CO for the available cells) during the initial growth of the culture or thereafter, the gas feed rate and/or agitation rate is reduced until the substrate inhibition is relieved.”
At paragraph 0085, Gaddy, et al., discuss supplying excess carbon monoxide and hydrogen. They state:
“A slight excess of CO and H2 is achieved by attaining steady operation and then gradually increasing the gas feed rate and/or agitation rate (10% or less increments) until the CO and H2 conversions just start to decline.”
For deep tank reactors, the height of the aqueous menstruum is a primary determinant of the contact time for the bioconversion to occur. This height also is a determinant of the static head at the bottom portion of the reactor. Higher pressures result in smaller bubble sizes and higher partial pressures both of which enhance mass transfer efficiency and gas substrate conversion efficiency in the fermentation reactor. Thus, on a commercial scale, deep tank reactors have a depth of at least about 10, preferably at least about 15, meters and use microbubbles of gas feed in order to achieve molar conversion efficiencies of at least about 60 percent of the total hydrogen and carbon monoxide supplied to the reactor. However, these operating parameters increase the risk of carbon monoxide inhibition.
The risk of carbon monoxide inhibition in deep tank reactors is more pronounced in start-up. Typically at start-up, the reactor is supplied with a culture of microorganisms from a seed farm and the size of the culture is limited, usually to about 10 percent or less of the concentration of the culture in the reactor at steady state. The reactor is then operated to obtain a robust growth of the culture to the sought density for steady state operation. The very dilute culture concentration is more subject to damage than would be the denser culture at steady state. Especially with commercial scale reactors, i.e., those with liquid capacities of at least 1 million, and more often at least about 5, say, 5 to 25, million, liters, sufficient culture volume to completely fill the capacity of the reactor is generally unavailable. Consequently, the reactor can be only partially filled with aqueous menstruum and the culture must be grown first to increase density in that portion filled. Then additional liquid is added to the reactor, and the culture must undergo further growth to viable densities. This process is continued until the deep tank reactor has reached its sought capacity.
To reduce the time required for start-up, it is desired to provide sufficient substrate to the microorganisms for robust growth. However, to reduce risks of overdosing the microorganisms with carbon monoxide, a common practice has been to undersupply gas substrate during start-up and thus extend the duration of the start-up process.
Processes are therefore sought to capture the benefits provided by a deep tank fermentation system at steady state conditions yet be able to rapidly start-up the deep, tank reactor without undue risk of damage to the microorganism culture.